Modular Evolution of Secretion Systems and Virulence Plasmids in Agrobacterium PDF

Chou et al. BMC Biology (2022) 20:16 https://doi.org/10.1186/s12915-021-01221-y RESEARCH ARTICLE Open Access Modular evolution of secretion systems and virulence...

Chou et al. BMC Biology https://doi.org/10.1186/s12915-021-01221-y RESEARCH ARTICLE Modular evolution of secretion systems and virulence plasmids in a bacterial species complex Lin Chou1†, Yu-Chen Lin1†, Mindia Haryono1, Chih-Feng Wu1,4, Jeff H. Chang4, Erh-Min Lai1,2,5 Abstract Background: Many named species as defined in Uncertainties regarding the organization of Agrobacterium tumefaciens species complex (a.k.a. phytopathogenicity and applications in transformation, genome diversity and evolution of species complexes. Results: We utilized 35 genome assemblies, balanced sampling of A. tumefaciens. Our genomic previously are distinct biological species Furthermore, our inference of gene content critical in fitness and ecology. For the type thought to be conserved, we detected multiple plasmids (pTi) and pTi-encoded type IV secretion phytopathogenicity, we uncovered novel diversity complex. Intriguingly, for both T6SS and T4SS, whereas extensive diversity exists for genes * Correspondence: [email protected] to this work. Sinica, Taipei, Taiwan Program, Taiwan International and Academia Sinica, at the end of the article Author(s). 2022 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if were made. The images or other third party material in this article are included in the article's Creative Commons unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the made available in this article, unless otherwise stated in a credit line to the data. Page 2 of 20 of a phylogeny-guided sampling scheme and an emphasis on approach for robust analysis in evolutionary genomics. We show effector binding and delivery can be classified into distinct types patterns of VgrG-associated domains and the neighboring genes be used to infer possible interacting partners. Similarly, the the pTi type among these strains can be used to infer phenotype- multi-level investigations at scales that range from whole genomes to between- to within-species are applicable to other bacteria. evolution of genes and domains is useful for inferring works. Virulence, Plasmid, Molecular evolution provide information such as gene content or phylogen- etic relationships, which are critical in understanding biological entities. Second, phylogenetic relationships among closely related bacterial strains often cannot be resolved with confidence, yet such information is funda- mental for evolutionary analysis. Third, comparative genomics studies are often limited by taxon sampling and/or assembly quality of available genome sequences. In this study, we utilized the Agrobacterium tumefaciens species complex, also known as Agrobacterium biovar 1 , as the study system for developing strategies that provide appropriate sampling and utilize multifaceted gen- omic analysis to resolve species boundaries and to investi- gate molecular evolution of key traits. These bacteria are known as the causative agents of crown gall disease that affects over 90 plant families. More importantly, the development of Agrobacterium-mediated transformation has provided a critical tool for genetic manipulation in plant sciences and agricultural biotechnology [14, 15]. Due to their importance, this complex has been studied for over a century and was found to harbor extensive phenotypic and genetic diversity that continues to con- found efforts to resolve their taxonomy [12–14]. Various methods, such as DNA-DNA hybridization, biochemical characteristics, and molecular markers have been used to define 10 genomospecies (i.e., G1-G9 and G13), which have continually been associated to new nomenclature, a process that causes greater confusion than resolution [5, 6, 16–18]. For example, the reference strain C58 used in many A. tumefaciens studies [15, 19, 20] belongs to G8, for which the name Agrobacterium fabrum was proposed in 2011. This has resulted in mixed usage of two names with different meanings (i.e., A. tumefaciens for the entire complex and A. fabrum for G8) in databases and lit- erature. Compounding confusion, the name Agrobacter- ium radiobacter refers to A. tumefaciens G4 [18, 21] and is also a heterotypic synonym of A. tumefaciens. Hereafter, we use A. tumefaciens in reference to the entire species complex and specific designations (i.e., G1–G9 and G13) for the genomospecies. Page 3 of 20 we identified 12 strains that represent six poorly charac- terized genomospecies (Table 1). Additionally, two Agro- bacterium larrymoorei strains, which represent the most closely-related sister lineage of A. tumefaciens [9, 18], were included as the outgroup. Whole-genome sequen- cing with substantial efforts in iterative improvements of the assemblies based on experimental and bioinformatic approaches were conducted for these 14 strains. Add- itionally, we selected 21 representatives from the 98 A. tumefaciens genome assemblies available from GenBank (Table 1) to yield a dataset with maximal genetic di- versity without emphasis on including pathogenic strains. To ensure balanced sampling, we selected be- tween two and five strains for each of the 10 recognized A. tumefaciens genomospecies. Importantly, 19 of these 35 assemblies, including nine produced in this study, are complete and most others are nearly complete (i.e., aver- age N50 = 1.3 Mb; cf. the two chromosomes are ~ 2.9 and ~ 2.3 Mb, respectively). Based on the homologous gene clustering results among these strains, we identified a core genome of 2093 single- copy genes, which correspond to ~ 40% of the genes anno- tated in each individual genome sequence. Compared to previous studies that conducted genome-based phylogen- etic analysis for Agrobacterium or higher taxonomic ranks [8, 9, 29], the more focused sampling in this study yielded a higher core gene count by one-to-two orders of magni- tude. This increase in core gene count and the improve- ment in taxon sampling allowed for the inference of a well-resolved maximum likelihood phylogeny of the A. tumefaciens species complex (Fig. 1A). Each of the 10 cur- rently recognized genomospecies forms a distinct mono- phyletic clade with > 80% bootstrap support. Additionally, we identified two novel genomospecies, G21 and G22, each represented by a single strain. The pattern of overall genome similarities exhibits a discrete multimodal distri- bution that supports use of a 95% ANI cutoff for delineat- ing bacterial species and quantifies the divergence of the A. tumefaciens complex from its most closely related sister lineage (Fig. 1B). These 12 A. tumefaciens genomospecies were classified into seven groups based on 90% ANI and further assigned to three supergroups according to the phyl- ogeny (Fig. 1A). Based on the time-calibrated phylogeny reported in Weisberg et al. , the most recent common ancestor (MRCA) of A. tumefaciens emerged ~ 48 mil- lion years ago (Mya) with a 95% highest posterior dens- ity (HPD) interval of 38.5–58.0 Mya, the three supergroups diverged ~ 40 Mya (95% HPD interval 31.5–48.1 Mya), and most of the recognized genomospe- cies emerged ~ 2–7 Mya (95% HPD interval 1.2–9.7 Mya). The large 95% HPD intervals suggest uncertainties regarding these time estimates. Regardless of the exact divergence time, the inferred rapid radiation in the early Page 4 of 20 this study. These include 14 new genomes derived from this study and 21 additional larrymoorei strains are included as the outgroup. Species name abbreviations: At, larrymoorei Coding sequences Pseudo-genes Geographic origin Isolation source 167 MD, USA Euonymus sp. 169 MT, USA Soil 145 CA, USA Achillea ptarmica 157 OR, USA Cranberry 218 ? Plant 233 France Human 223 France Human 146 Belgium Ditch water 180 France Antiseptic flask 148 France Human 199 Tunisia Prunus dulcis 207 CA, USA Juglans regia 167 ? ? 269 CA, USA Rubus sp. 198 France Food 199 France Human 135 Harbin, China Soil 240 South Africa Dahlia sp. 236 Israel Dahlia sp. 254 CA, USA Medicago sativa 213 UK Flacourtia ramontchi 285 Tunisia Pyrus communis 178 ? ? 169 CA, USA Cerasus pseudocerasus 811 China Plant 29 NY, USA Cerasus pseudocerasus 157 Australia Soil 253 Australia Soil 203 Belgium Solanum lycopersicum 94 France Rhizospheric soil from Prunus persicae 167 ? ? 162 Japan Cucumis melo 162 FL, USA Bos taurus feces 142 FL, USA Ficus benjamina 95 Italy ? framework for our downstream examination of gene and domain phylogenies. The high levels of assembly completeness provided confident inference of gene content comparisons. The principal coordinate analysis and hierarchical clustering results indicated that all 12 A. tumefaciens genomospecies Page 5 of 20 the Agrobacterium tumefaciens species complex. The sister species Agrobacterium larrymoorei (A. l.) phylogeny based on a concatenated alignment of 2093 single-copy genes shared by all 35 Bootstrap support values in the range of 60–80% are labeled. Strains with complete genome (“*”). The genomospecies assignments (i.e., G1-G9, G13, and G21-G22) are labeled to the right of strain are indicated by the colored background of the genomospecies assignments. Information to the the grouping of genomes according to different cutoff values of genome-wide average of type VI secretion system (T6SS)-encoding gene cluster (green: present; white: absent), copy of plasmids, and the tumor-inducing plasmid (pTi) type based on k-mer profile (white based on the percentages of genomic segments mapped and the ANI values Page 6 of 20 state in the MRCA of the A. tumefaciens complex, pres- ence of the T6SS genes appears to be a more parsimoni- ous hypothesis based on the presence of these genes in the outgroup (Fig. 2). However, the lack of synteny con- servation between the linear chromosomes of A. tumefa- ciens and A. larrymoorei and the variable locations of vgrG homologs (Additional file 1: Figure S2) suggest that multiple independent origins are also possible. For broader scales, the T6SS genes have a patchy distribu- tion among Rhizobiaceae [33, 36], indicating that these genes are not essential for these bacteria and have high rates of gains and losses. Consistent with this, multiple pseudogenes confirmed by manual curation of annota- tion were found (e.g., tssH in S56 and tssA/vgrG in ATCC31749) (Fig. 2), suggesting that for some strains these gene clusters are in the process of degradation and will be eventually lost. Examination of synteny revealed that the imp operon, which encodes the majority of T6SS structural compo- nents , is more conserved in gene composition and order than the hcp operon, which often has different genes downstream of vgrG (Fig. 2). This genetic diversity may play a key role in interbacterial competition because genes downstream of vgrG include those that encode ef- fector and immunity (EI) protein pairs. The agrobac- terial T6SS effectors often correspond to different toxins, and the cognate immunity proteins provide pro- tection against self-intoxication [7, 27, 33, 36]. The rapid evolution of EI gene pairs is illustrated by three exam- ples. First, despite the low levels of sequence divergence among G1 strains (i.e., > 98% ANI in all pairwise com- parisons), different genes are found downstream of their vgrG homologs and this variation is not consistent with either the species phylogeny or the T6SS core gene phyl- ogeny. Second, strain CFBP4996 of G7 has homologs of the same EI gene pair as strains 12D1 and 183 of G4, ra- ther than with other members of G7. Third, in both G3 strains, vgrG and the associated EI genes are located elsewhere on the linear chromosome, rather than being a part of the hcp operon (Fig. 2 and Additional file 1: Figure S2). These results suggest that recombination in- volving gene modules has contributed to the genetic di- versity of these A. tumefaciens T6SS EI gene pairs. Modularity of VgrG and its associated EI pair The knowledge that vgrG homologs encode proteins with distinct C-terminal domains responsible for binding speci- ficities of different T6SS effectors for delivery suggested that each vgrG homolog and its downstream EI gene pair may evolve as a functional module [33, 36]. Here, we sought to investigate the patterns of gene co-occurrence and intra-module recombination to better understand the diversity and evolution of these genes. For in-depth inves- tigation of vgrG evolution, we began by examining domain Page 7 of 20 gene cluster. The maximum likelihood phylogeny was inferred based on a concatenated 14 core T6SS genes, including tagE, tagF, tssM, tssL, tssK, tagH, tssG, tssF, tssE, tagJ, tssC40, tssC41, tssB, are excluded because some homologs are pseudogenized. Genes downstream of tssD (e.g., tai, presence. The species phylogeny on the right is based on Fig. 1. Genes are color-coded according to by gray blocks Intriguingly, despite conservation of domain architecture within each subtype (Fig. 3), the phylogenies inferred from the three domains encoded by all vgrG homologs do not have the same topology (Additional file 1: Figure S5). For domain 1, sequences from the same subtype do not always form monophyletic groups. For domains 2 and 3, the short sequence lengths limited the phylogenetic resolution; none- theless, low divergence within the same subtype and high di- vergence between subtypes were observed. These patterns suggest that each domain-encoding region evolved inde- pendently and can recombine between subtypes. At the level of gene cluster organization, vgrG homo- logs within a subtype can have distinct downstream genes (e.g., A1 and B1), regardless of whether they are associated with the main T6SS gene cluster (Fig. 4 and Additional file 2: Dataset S1). These findings suggest that in addition to domain shuffling among vgrG homo- logs , recombination also facilitated novel vgrG-ef- fector pairings in the evolution of these T6SS genes. When T6SS diversity was examined in a phylogenetic context, numbers, and types of vgrG homologs, as well as their linked EI genes, lack strong correlations with species phylogeny (Fig. 5). Based on our manual curation of vgrG-associated genes, a total of 63 putative effector genes were identified (Additional file 2: Dataset S2). Among these, peptidoglycan-targeting toxins and nucle- ases are the two most commonly found categories with Page 8 of 20 of vgrG homologs. Six major types with two types having subtypes (A1-3 and D1-2) within them each homolog, the genomospecies assignment is provided in square brackets, followed by the for those functionally characterized homologs (i.e., vgrG1-2 for C58 homologs and vgrGa-d Page 9 of 20 The grouping and labeling of vgrG homologs are based on the convention used in Fig. 3. For plotted to illustrate if it is associated with the main T6SS gene cluster or not, and 10 downstream genes. Two A1-type homologs from the strain S56 (i.e., AGR1B_RS27940 and each other and plotted together et al. , EI4/11 in Wu et al. , EI7 in Santos et al. , and EI2/5/8/9/10 in Wu et al.. The virulence plasmids and associated genes The tumor-inducing plasmids (pTi) are an important component of A. tumefaciens genomes. These large accessory replicons harbor the virulence (vir) regulon genes that encode the Vir proteins and type IV secretion system (T4SS) for processing and delivering a transfer Page 10 of 20 and T6SS effector/immunity genes. The species tree is based on Fig. 1. Gene presence is gene copy numbers are labeled when applicable Figure S6), and sequence divergence of core genes (Add- itional file 1: Figure S7). Moreover, their T-DNA regions also differ from the typical sizes of ~ 18–26 kb observed in types I–III pTi (Fig. 6). For the tumorigenic strain CFBP5473 (Additional file 1: Figure S8), the predicted T-DNA border sequences flank an exceptionally large (~ 93 kb) region that contains all of the vir regulon genes in addition to the typical T-DNA-associated genes (e.g., synthesis of opine and plant hormone) (Fig. 6). This re- flects either a translocation of a T-DNA border sequence or reliance on a non-canonical T-DNA border sequence that we could not identify. For pTiCFBP4996, its 7-kb T-DNA is predicted to contain only four genes (i.e., two correspond to opine synthesis and two encode hypothet- ical proteins). Plant hormone synthesis genes, which are necessary to cause visible disease symptoms, were not identified in this predicted T-DNA region or elsewhere on this plasmid. Consistent with predictions, strain CFBP4996 did not induce tumor formation when inocu- lated onto stems of tomato plants (Additional file 1: Page 11 of 20 include 15 from the genome data set listed in Table 1 and five additional pTi type assignments are based on k-mer profiles. A type V pTi from Agrobacterium abbreviations: At, Agrobacterium tumefaciens; Al, Agrobacterium larrymoorei; Av, Accession Size (bp) Coding sequences Geographic origin Isolation source NZ_CP032925 176,213 159 MD, USA Euonymus sp. NZ_CP011249 194,264 153 CA, USA Achillea ptarmica NZ_CP029048 192,674 173 Tunisia Prunus dulcis NZ_CP042277 177,577 159 CA, USA Juglans regia NZ_CP032929 214,233 198 CA, USA Rubus sp. NC_002147 206,479 195 Japan Cerasus pseudocerasus NZ_CP039893 220,025 178 South Africa Dahlia sp. NZ_CP039902 220,025 181 Israel Dahlia sp. NZ_CP026926 166,117 138 CA, USA Medicago sativa NZ_CM016551 605,495 521 UK Flacourtia ramontchi NZ_CP039927 189,955 176 Tunisia Pyrus communis NZ_CP033026 177,577 160 CA, USA Cerasus pseudocerasus NC_003065 214,233 197 NY, USA Cerasus pseudocerasus NZ_CM016546 248,634 225 Australia Soil NC_010929 244,978 223 Germany Dahlia sp. KX388536 197,268 210 FL, USA Chrysanthemum sp. Succinamopine KX388535 176,375 194 CT, USA Euonymus sp. NZ_CP039694 404,101 360 FL, USA Ficus benjamina NZ_CM016547 190,036 169 Italy ? NC_011982 258,824 193 Hungary Vitis sp. among pTi types were observed (Fig. 7). For example, while all pTi harbor a conserved virE3a that facilitates T-DNA protection and entry into host , type I pTi harbor one or two additional copies of virE3 that belong to different sequence types (i.e., sharing the same anno- tation but classified as distinct homologs due to se- quence divergence). Similarly, virF, which encodes an F- box protein that is a putative host-range determinant , can be classified into three sequence types with dis- tinct distributions. Those less well-characterized vir genes, such as virD3 [46, 47], virJ , and virP, are also distributed differently among pTi types. Finally, in addition to the presence/absence of individual genes, the overall organization of vir regulons also differ among these pTi (Additional file 1: Figure S9). All type I pTi are conserved in sharing a ~ 40-kb region that contains all vir genes. In comparison, locations of vir genes are more variable among type II/III pTi; virF/P (and virQ/H if present) are located ~ 5–50 kb away from the main vir gene cluster. Other than vir genes, the gene content and. For the other vir regulon genes, several differences organization of T-DNA are also different (Fig. 8). All Page 12 of 20 of pTi. The maximum likelihood phylogeny was inferred based on the concatenated alignment sites (Supplementary Figure S6A). The species phylogeny on the right is based on Fig. 1. The pTi in bold. For those with relevant information available, the genomospecies assignments are are indicated in parentheses. For the alignment, all plasmids are visualized in linear form starting are color-coded according to functions, and the predicted T-DNA regions are indicated by the Syntenic regions are indicated by gray blocks [8, 10]. Regardless of the exact mechanisms, these pat- terns supported application of the biological species con- cept, which is based on genetic barriers, to these A. tumefaciens genomospecies and other bacteria. With the continuing drop in sequencing cost, ANI analysis can serve as a standard approach for accurate classifica- tion of additional strains , which in turn could facili- tate research and communication, and ideally leads to improvements in bacterial taxonomy for basic works and applications. It is worth noting that despite our effort, the 33 strains included in this study do not fully capture the diversity of the A. tumefaciens species complex. For example, the species Agrobacterium arsenijevicii , Agrobacterium nepotum (G14) , Agrobacterium viscosum (G15) , and two unnamed genomospecies G19/G20 are also members of this species complex. As more strains are characterized in the future, it is likely that higher levels of diversity within this group will continue to be discov- ered. Furthermore, several potentially confusing issues regarding Agrobacterium taxonomy remain to be re- solved. For example, strain MAFF210266 that we re- ferred to as a representative of G21 shares 98% ANI with an important strain K599 (=NCPPB2659). However, K599 harbors a root-inducing plasmid (pRi) associated with hairy root disease and was named as Page 13 of 20 putative pTi. The pTi sequences derived from this study are highlighted in bold. For those with assignments are indicated in square brackets and the opine types are indicated in parentheses. and empty circles, respectively. The main vir genes include the components of the type IV For virE3, the three sequence types are listed separately. For virF and virD3, the sequence types are genes virA and virJ are located on another plasmid and plotted as absent in this figure. The locus tags could be due to biases in the sampling of available ge- nomes, or the all-against-all pairwise comparisons in- cluded mostly distantly-related species. In our study that provided a detailed examination of closely related species, the ~ 85–93% ANI among A. tumefaciens geno- mospecies (Fig. 1B) indicated that the A. tumefaciens complex is indeed a coherent entity with high divergence from its closest sister lineage within the same genus (i.e., A. larrymoorei). The driving forces for maintaining spe- cies complexes and the prevalence of such above-species level entities are interesting questions that require fur- ther investigations. For the A. tumefaciens complex, al- though the nomenclature originated from its phytopathogenicity, it is well-established that this group contains both pathogenic and non-pathogenic strains that differ in the possession of an oncogenic plasmid (i.e., pTi) or not. The promiscuous nature of their pTi [9, 56–58] (Fig. 6) suggested that lineages within this complex may experience frequent transitions between Page 14 of 20 on tumor-inducing plasmids (pTi). Two unusual putative pTi sequences (i.e., pTiCFBP4996 and derived from this study are highlighted in bold. For those with relevant information indicated in square brackets and the opine types are indicated in parentheses. Genes are color- are indicated by gray blocks taxonomy. Such situations demonstrated the chal- lenges of establishing a new standardized taxonomy even when the practical issues of transitioning from the current taxonomy are not considered, and perhaps is to be expected given the highly variable evolutionary rates across different lineages. Given these considerations, other aspects of biology (e.g., physiology, ecology) may play more important roles in defining those higher taxo- nomic ranks. classifica- Units and modularity of molecular evolution For evolutionary studies, the levels at which selection and other processes operate on have been a topic that received much attention. For prokaryotes, levels from the entire genome to individual functional domains within genes are of particular interest. Based on our re- sults, all of these levels must be considered to compre- hend the complex patterns. At the whole-genome level, the clear species boundar-. However, while it may be desirable to establish a ies based on overall similarity (Fig. 1 and Additional file 1: Figure S1) suggested that the entire genome largely evolves as a single coherent unit. This result is consist- ent with previous findings that at the global level HGT has very little impact on the reconstruction of organis- mal phylogeny [64, 65], despite the extensive HGT in- ferred in bacterial evolution [65–68] and the importance of HGT in adaptation [69–71]. A possible explanation for these seemingly conflicting observations is that most Page 15 of 20 that may reflect functional constraints were observed at finer scales. For example, each vgrG is linked to its cognate ef- fector/chaperone genes, and each effector is linked to its cognate immunity gene. Such modularity is expected to be maintained by selection, similar to the observations regard- ing co-transfers of genes involved in associated biochemical pathways. However, the linkage could be broken down by recombination at within- or between-species levels, as evident in the diversity of vgrG gene neighborhoods, even for those homologs belonging to the same type (Fig. 4). Fourth, at the level of individual genes, the within-genome diversity of vgrG homologs (Fig. 5) provides further support to the hypothesis that HGT is more important than duplica- tions in driving gene family expansions in bacteria. Fi- nally, at the intra-gene level, within- or between-species recombination may be important in generating novel com- binations of domains, thus promoting the diversification of homologs (Fig. 3 and Additional file 1: Figure S6). Taken together, these observations illustrated the complexity of biological systems. While it is difficult to draw up generalized rules or to estimate the relative im- portance of each evolutionary process at different levels, it is important to consider and examine these complex- ities to better understand organisms of interest. Conclusions In summary, by using a group of important bacteria as the study system, this work utilized a strategy of phylogeny- conscious genome sampling for systematic investigations of a species complex. This approach requires prior know- ledge regarding the extant phylogenetic diversity of the study system, which is important in the planning stage of large-scale genome-sequencing projects. In addition to improving the cost-effectiveness, this strategy is also crit- ical in obtaining an unbiased picture that does not over- emphasize certain subgroups. Furthermore, the emphasis on generating and utilizing high-quality assemblies im- proves the confidence in gene content analysis. With the continuing advancements in sequencing technologies and bioinformatic tools, this emphasis becomes increasingly accessible. For this study system, our examination of bio- logical boundaries at the species level and above improves the understanding of how natural biodiversity is orga- nized. The targeted analysis of those secretion system genes and oncogenic plasmids provides novel insights re- garding the key genetic variations involved in the fitness and ecology of these soil-borne phytopathogens that need to compete in complex microbiota and invade plant hosts. Moreover, the multi-level analysis of their genetic diversity from whole-genome to intra-genic domains highlights the complexity of these biological systems. The strategy and findings of this work provide useful guides for future stud- ies of other bacteria. Page 16 of 20 v2.6.0 with e-value cutoff set to 1e−15 and OrthoMCL v1.3 were used to infer the homologous gene clusters. The result was converted into a matrix of 35 genomes by 17,058 clusters, with the value in each cell corresponding to the copy number. This matrix was further converted into a Jaccard distance matrix among genomes using the VEGAN package v2.5-6 in R, then processed using the principal coordinates analysis func- tion in the APE package and visualized using ggplot2 v3.3.2. The hierarchical clustering analysis was performed using PVCLUST v3.4.4. For phylogenetic analysis, homologous sequences were aligned using MUSCLE v3.8.31 for maximum likeli- hood inference by PhyML v.3.3.20180214. The pro- portion of invariable sites and the gamma distribution parameter were estimated from the data set, and the number of substitute rate categories was set to four. The bootstrap supports were estimated based on 1000 replicates. Analysis of the type VI secretion system genes To identify the T6SS-associated genes, C58 [27, 36, 37] and other strains [7, 33] that have been characterized ex- perimentally were used as the references. Based on the known T6SS genes in these genomes, homologous genes. Due to the difficulties of identifying appropriate in other genomes were identified based on the OrthoMCL result. To screen for novel T6SS effector, chaperone, and immunity genes, genes that are located near vgrG (i.e., three upstream and ten downstream) were examined manually by using the NCBI conserved domain database (CDD) and the Phyre2 protein fold recognition server ; the e-value cutoff was set to 0.01. A few genes with a hit to the DUF4123 domain (i.e., a known T6SS chaperone) but have an e-value above the cutoff were manually added back to the list of T6SS-associated genes (e.g., CFBP5477_RS20350). To confirm the absence of specific T6SS genes, the genome sequences were used as the subjects and the protein se- quences of known genes were used as the queries to run TBLASTN searches. For classification of the vgrG homologs, we developed a domain-based scheme. The conserved N-terminal TIGR03361 domain was first identified by the NCBI CDD searches. A global alignment of all homologs was used to determine the exact boundaries of this domain. After this TIGR03361 domain was removed, the remaining C-terminal sequences were processed using MEME v5.1.1 to identify conserved domains that meet these criteria: (1) present in at least two sequences, (2) with zero or one occurrence per sequence, (3) with a size between 30 and 300 a.a., and (4) with an e-value lower than 0.05. The results were manually curated to break down large domains that are composed of smaller domains. Pairwise BLASTP searches were conducted to Page 17 of 20 week-old seedlings. Bacterial strains were transferred from stock to 5 mL 523 broth and cultured overnight at 28 °C in a shaker incubator (250 rpm), then sub-cultured for 4 h prior to inoculation. Bacterial cells were washed and re- suspended in 0.9% NaCl solution with a concentration of OD600 0.2. The stem was punctured with a sterilized sew- ing needle, and 5 μL of bacterial suspension was added to the wounding site. The plants were collected 3 weeks after inoculation and 1-cm stem segments centered at the wounding site were cut for weighing. Supplementary Information The online version contains supplementary material available at https://doi. org/10.1186/s12915-021-01221-y. Additional file 1: Figure S1. Gene content dissimilarity among the Agrobacterium genomes. (A) and (B): principal coordinate analysis with and without the outgroup A. larrymoorei, respectively. The % variance explained by each axis is provided in parentheses. (C) and (D): hierarchical clustering with and without the outgroup A. larrymoorei, respectively. Figure S2. Global alignment of the linear chromosomes. Locations of T6SS-hcp operons and vgrG homologs are labeled. Figure S3. Logo plots of the putative protein domains identified among vgrG homologs. For each domain, the length and the number of homologs with the domain is labeled. Domain 1 is the only domain with a corresponding database entry (TIGR03361). Figure S4. Predicted structures of VgrG homologs. Regions are colored according to the scheme used in the domain analysis (Fig. 3). The chain D of PA0091 VgrG1 (PDB identifier: 4MTK) from Pseudomonas aeruginosa was selected as the template. The C-terminal parts that could not be confidently in- ferred are omitted. In all cases, the coverage (i.e., percentage of the se- quence included in the structure prediction) are at least 75%, the sequence identity to the template is at least 30% and the confidence score is 100%. Figure S5. Maximum likelihood phylogenies of vgrG-asso- ciated domains. (A) Domain 1 (TIGR03361; VI_Rhs_Vgr super family), (B) Domain 2 (unknown function), and (C) Domain 3 (unknown function). Figure S6. Principal coordinate analysis of gene content among the pu- tative pTi analyzed. Figure S7. Maximum likelihood phylogeny of pTi based on the concatenated alignment of shared single-copy genes. (A) All of the 20 pTi sequences analyzed; 21 core genes and 8534 aligned amino acid sites. (B) Excluding the two novel pTi; 40 core genes and 15,473 aligned amino acid sites, all branches received > 80% bootstrap support. Figure S8. Tomato tumor assay of strains 12D1, CFBP4996, and CFBP5473. Mock was inoculated with sterilized water as a negative con- trol and strain C58 was included as a positive control. Strain 12D1 harbors a plasmid with opine transporter and catabolism genes but lacks vir regu- lon genes and identifiable T-DNA. CFBP4996 and CFBP5473 harbor novel types of putative tumor-inducing plasmids (pTi). (A) Tomato stems at three weeks after inoculation. Scale bar: 0.25 cm. (B) Weight distribution of five biological replicates (1-cm segments of the stem centered at the inoculation site). The letters indicate ANOVA results. Figure S9. Gene organization of the vir regulons on pTi. Syntenic regions are indicated by grey blocks. The virulence (vir) genes are highlighted in red, the conjuga- tion (tra) genes are highlighted in yellow, and other genes are plotted in white. Additional file 2: Dataset S1. List of vgrG-associated genes. Information including genomic location, RefSeq annotation, and domain prediction are included. Dataset S2. Locus tags of the vir regulon genes on pTi. The virA/J of 1D1609 are located on another plasmid and are highlighted by “*”. The virB7 of pTiChry5 and pTiEU6 are unannotated in the GenBank RefSeq records so no locus tag is available but the gene presence was confirmed by BLASTN searches. plants (culti- Acknowledgements We thank Ai-Ping Chen, Hsin-Ying Chiang, Shu-Jen Chou, Mei-Jane Fang, Ya- Yi Huang, Wen-Sui Lo, and Javier F. Tabima for technical assistance. Sophien Page 18 of 20 5. Popoff MY, Kersters K, Kiredjian M, Miras I, Coynault C. Position taxonomique under the permits 103-B-003 de souches de Agrobacterium d’origine hospitalière. Ann Inst Pasteur of Taiwan. The Sanger se- Microbiol. 1984;135:427–42. library preparation service 6. Costechareyre D, Bertolla F, Nesme X. Homologous recombination in (Institute of Plant and Mi- Agrobacterium: potential implications for the genomic species concept in MiSeq sequencing service was bacteria. Mol Biol Evol. 2009;26:167–76. Biology, Academia 7. Wu C-F, Santos MNM, Cho S-T, Chang H-H, Tsai Y-M, Smith DA, et al. Plant- service was provided by pathogenic Agrobacterium tumefaciens strains have diverse type VI effector- City, Taiwan). The Institute of immunity pairs and vary in in-planta competitiveness. Mol Plant Microbe and the Department of Botany Interact. 2019;32:961–71. 8. Lassalle F, Planel R, Penel S, Chapulliot D, Barbe V, Dubost A, et al. Ancestral genome estimation reveals the history of ecological diversification in Agrobacterium. Genome Biol Evol. 2017;9:3413–31. 9. Weisberg AJ, Davis EW, Tabima J, Belcher MS, Miller M, Kuo C-H, et al. JHC, EML, CHK. Investigation: Unexpected conservation and global transmission of agrobacterial virulence YCL, MNS, STC, CHK. Project plasmids. Science. 2020;368:eaba5256. Validation: LC, YCL, MH, 10. Jain C, Rodriguez-R LM, Phillippy AM, Konstantinidis KT, Aluru S. High draft: LC, YCL, CHK. throughput ANI analysis of 90 K prokaryotic genomes reveals clear species JHC, EML, CHK. All authors boundaries. Nat Commun. 2018;9:5114. 11. Murray CS, Gao Y, Wu M. Re-evaluating the evidence for a universal genetic boundary among microbial species. Nat Commun. 2021;12:4059. 12. Young JM. Agrobacterium—taxonomy of plant-pathogenic Rhizobium species. In: Tzfira T, Citovsky V, editors. Agrobacterium Biol Biotechnol. New National Institute of Food York: Springer; 2008. p. 183–220. Available from: http://link.springer.com/ awards 2014-51181-22384 chapter/10.1007/978-0-387-72290-0_5. was supported by Academia 13. Kado CI. Historical account on gaining insights on the mechanism of crown of Taiwan (MOST 104- gall tumorigenesis induced by Agrobacterium tumefaciens. Front Microbiol. Research in the Kuo lab 2014;5:340. of Science and Technol- 14. Nester EW. Agrobacterium: nature’s genetic engineer. Front Plant Sci. 2015;5: The fun- 730. and interpretation, or the 15. Hwang H-H, Yu M, Lai E-M. Agrobacterium-mediated plant transformation: biology and applications. Arab Book. 2017;15:e0186. 16. Mougel C, Thioulouse J, Perrière G, Nesme X. A mathematical method for determining genome divergence and species delineation using AFLP. Int J Syst Evol Microbiol. 2002;52:573–86. 17. Portier P, Saux MF-L, Mougel C, Lerondelle C, Chapulliot D, Thioulouse J, et al. Identification of genomic species in Agrobacterium biovar 1 by AFLP genomic markers. Appl Environ Microbiol. 2006;72:7123–31. 18. Costechareyre D, Rhouma A, Lavire C, Portier P, Chapulliot D, Bertolla F, et al. Rapid and efficient identification of Agrobacterium species by recA allele analysis: Agrobacterium recA diversity. Microb Ecol. 2010;60:862–72. 19. Hellens R, Mullineaux P, Klee H. A guide to Agrobacterium binary Ti vectors. Trends Plant Sci. 2000;5:446–51. 20. Lee L-Y, Gelvin SB. T-DNA binary vectors and systems. Plant Physiol. 2008; 146:325–32. 21. Lassalle F, Campillo T, Vial L, Baude J, Costechareyre D, Chapulliot D, et al. Genomic species are ecological species as revealed by comparative genomics in Agrobacterium tumefaciens. Genome Biol Evol. 2011;3:762–81. 22. Young JM, Pennycook SR, Watson DRW. Proposal that Agrobacterium radiobacter has priority over Agrobacterium tumefaciens. Request for an Sinica, Taipei, Taiwan. Opinion. Int J Syst Evol Microbiol. 2006;56:491–3. Program, Taiwan International 23. Slater SC, Goldman BS, Goodner B, Setubal JC, Farrand SK, Nester EW, et al. and Academia Sinica, Genome sequences of three Agrobacterium biovars help elucidate the evolution of National Chung Hsing multichromosome genomes in bacteria. J Bacteriol. 2009;191:2501–11. Botany and Plant Pathology, 24. Goodner B, Hinkle G, Gattung S, Miller N, Blanchard M, Qurollo B, et al. 5Biotechnology Center, Genome sequence of the plant pathogen and biotechnology agent Agrobacterium tumefaciens C58. Science. 2001;294:2323–8. 25. Wood DW, Setubal JC, Kaul R, Monks DE, Kitajima JP, Okura VK, et al. The genome of the natural genetic engineer Agrobacterium tumefaciens C58. Science. 2001;294:2317–23. 26. Haryono M, Cho S-T, Fang M-J, Chen A-P, Chou S-J, Lai E-M, et al. Differentiations in gene content and expression response to virulence for prokaryotes. FEMS induction between two Agrobacterium strains. Front Microbiol. 2019;10:1554. 27. Ma L-S, Hachani A, Lin J-S, Filloux A, Lai E-M. Agrobacterium tumefaciens WP. The bacterial species deploys a superfamily of type VI secretion DNase effectors as weapons for diversity. Science. 2009; interbacterial competition in planta. Cell Host Microbe. 2014;16:94–104. 28. Benson DA, Cavanaugh M, Clark K, Karsch-Mizrachi I, Ostell J, Pruitt KD, et al. universal across life’s domains. GenBank. Nucleic Acids Res. 2018;46:D41–7. 29. Ormeño-Orrillo E, Servín-Garcidueñas LE, Rogel MA, González V, Peralta H, bacterial species definition in Mora J, et al. Taxonomy of rhizobia and agrobacteria from the Rhizobiaceae Sci. 2006;361:1929–40. family in light of genomics. Syst Appl Microbiol. 2015;38:287–91. Page 19 of 20 SJ. Type VI secretion 53. Mafakheri H, Taghavi SM, Puławska J, de Lajudie P, Lassalle F, Osdaghi E. Two for bacterial competitiveness. novel genomospecies in the Agrobacterium tumefaciens species complex associated with rose crown gall. Phytopathology. 2019;109:1859–68. and diversity of type VI secretion 54. Valdes Franco JA, Collier R, Wang Y, Huo N, Gu Y, Thilmony R, et al. Draft genome sequence of Agrobacterium rhizogenes strain NCPPB2659. Genome Comstock LE, Basler M, et al. The Announc. 2016;4:e00746-16. as a disintegration weapon. PLoS 55. Singh NK, Lavire C, Nesme J, Vial L, Nesme X, Mason CE, et al. Comparative genomics of novel Agrobacterium G3 strains isolated from the International C-H, Lai E-M. Redundancy Space Station and description of Agrobacterium tomkonis sp. nov. Front loci in antibacterial activity of Microbiol. 2021;12:3369. Front Microbiol. 2020;10:3004. 56. Hooykaas PJJ, Klapwijk PM, Nuti MP, Schilperoort RA, Rörsch A. Transfer of Papadopoulos J, Bealer K, et al. the Agrobacterium tumefaciens Ti plasmid to avirulent agrobacteria and to Bioinformatics. 2009;10:421. Rhizobium ex planta. J Gen Microbiol. 1977;98:477–84. E-M. Secretome analysis 57. Haryono M, Tsai Y-M, Lin C-T, Huang F-C, Ye Y-C, Deng W-L, et al. Presence a type VI secretion system in of an Agrobacterium-type tumor-inducing plasmid in Neorhizobium sp. NCHU2750 and the link to phytopathogenicity. Genome Biol Evol. 2018;10: E-M. VgrG C terminus confers the 3188–95. is required for binding with PAAR 58. Rathore DS, Mullins E. Alternative non-Agrobacterium based methods for plant Acad Sci. 2016;113:E3931–40. transformation. In: Roberts JA, editor. Annu Plant Rev Online. Hoboken, New of the Agrobacterium type VI Jersey: John Wiley & Sons, Ltd.; 2018. p. 891–908. Available from: https:// onlinelibrary.wiley.com/doi/abs/10.1002/9781119312994.apr0659. 59. Barco RA, Garrity GM, Scott JJ, Amend JP, Nealson KH, Emerson D. A genus Mekalanos JJ. Type VI secretion definition for Bacteria and Archaea based on a standard genome protein into target cells where it relatedness index. mBio. 2020;11:e02475-19. 60. Parks DH, Chuvochina M, Waite DW, Rinke C, Skarshewski A, Chaumeil P-A, JB, Sauder JM, Pukatzki S, et al. A standardized bacterial taxonomy based on genome phylogeny tail-associated protein substantially revises the tree of life. Nat Biotechnol. 2018;36:996–1004. Proc Natl Acad Sci. 2009; 61. Parks DH, Chuvochina M, Chaumeil P-A, Rinke C, Mussig AJ, Hugenholtz P. A complete domain-to-species taxonomy for Bacteria and Archaea. Nat Biotechnol. 2020;38:1079–86. system in agrobacteria. mBio. 2021; 62. Kuo C-H, Ochman H. Inferring clocks when lacking rocks: the variable rates of molecular evolution in bacteria. Biol Direct. 2009;4:35. L, Dong TG. Identification of 63. Okasha S. Evolution and the Levels of Selection. Oxford: Oxford University a conserved chaperone domain. Press; 2006. Available from: https://oxford.universitypressscholarship.com/ view/10.1093/acprof:oso/9780199267972.001.0001/acprof-9780199267972 Wilton A, Diaz-Satizabal L, Pukatzki 64. Daubin V, Moran NA, Ochman H. Phylogenetics and the cohesion of diverse type VI secretion system bacterial genomes. Science. 2003;301:829–32. 65. Choi I-G, Kim S-H. Global extent of horizontal gene transfer. Proc Natl Acad NJ, Chang JH. Diversification of Sci. 2007;104:4489–94. bacteria. Philos Trans 66. Ochman H, Lawrence JG, Groisman EA. Lateral gene transfer and the nature of bacterial innovation. Nature. 2000;405:299–304. anchorage protein VirE3 for 67. Dagan T, Artzy-Randrup Y, Martin W. Modular networks and cumulative sites for T-DNA protection. Cell impact of lateral transfer in prokaryote genome evolution. Proc Natl Acad Sci. 2008;105:10039–44. host-range-determining virulence 68. Chan CX, Beiko RG, Darling AE, Ragan MA. Lateral transfer of genes and T-DNA transfer to Zea mays. Proc gene fragments in prokaryotes. Genome Biol Evol. 2009;1:429–38. 69. Pál C, Papp B, Lercher MJ. Adaptive evolution of bacterial metabolic virD3 gene is not essential networks by horizontal gene transfer. Nat Genet. 2005;37:1372–5. 70. Kuo C-H, Ochman H. The fate of new bacterial genes. FEMS Microbiol Rev. for virulence while virD3 and 2009;33:38–43. tumefaciens. Mol 71. Wiedenbeck J, Cohan FM. Origins of bacterial diversity through horizontal genetic transfer and adaptation to new ecological niches. FEMS Microbiol Rev. 2011;35:957–76. a Ti plasmid gene or a 72. Mira A, Ochman H, Moran NA. Deletional bias and the evolution of bacterial into plants. Mol Microbiol. genomes. Trends Genet. 2001;17:589–96. 73. Kuo C-H, Ochman H. Deletional bias across the three domains of life. K-C, Kado CI. Characterization Genome Biol Evol. 2009;1:145–52. tumefaciens strains and 74. Sundin GW. Genomic insights into the contribution of phytopathogenic assays. Plant Pathol. bacterial plasmids to the evolutionary history of their hosts. Annu Rev Phytopat

Modular Evolution of Secretion Systems and Virulence Plasmids in Agrobacterium PDF

Document Details

Tags

Related

Summary

Full Transcript